Abstract: Calculating a score for a chain of interactions in a call center may include: during a first training phase, train a first model which, given an interaction and interaction metadata, predict an initial estimated customer satisfaction score; during a second training phase, train a second model which, given an interaction and interaction metadata, text and metadata of an immediately preceding interaction in a chain of interactions, and features of the chain, predict a refined estimated customer satisfaction score; and during an inference phase: given a chain of interactions and metadata of each interaction, compute an initial estimated customer satisfaction score for each interaction using the first model; beginning with a second interaction in the chain and metadata of each interaction, compute a refined estimated customer satisfaction score for each interaction using the second model; combine the interaction scores into a combined customer satisfaction score; and output the score.

Abstract: A system and method for creating input data to be used to train a conversational bot may include receiving a set of conversations, each conversation including sentences, classifying each sentence into a dialog act taken from a number of dialog acts, for each set of sentences classified into a dialog act, clustering the set of sentences into clusters based on the content (e.g. text) of the sentences, each cluster having a cluster name or label, and generating a language model based on the cluster labels. Slots may be identified in the sentences based in part on the dialog act classifications. A bot may be trained using data such as the slots, language model, and clusters.

Abstract: A method, system, and non-transitory processor-readable storage medium for automatic key phrase rule generation for automatic key phrase extraction including: receiving a corpus sample including a plurality of documents containing text, receiving a plurality of identified key phrases which relate to a topic of the text of at least one corresponding document; assigning a part-of-speech to each word in the corpus sample; generating a part-of-speech pattern from each identified key phrase; and generating key phrase rules.

Abstract: A system and method for segmenting or dividing a series of computer-based actions, for example into sentences, may provide a sequence of subsets of the series of actions to a neural network using a sliding window, and divide or segment the series actions into segments at points where the loss of the neural network is above a threshold. The dividing may include, for each of a sequence of computer-based actions within a sliding window determining if the sequence when provided to the neural network corresponds to a loss above or equal to a threshold, and if so, determining that an action in the sequence of actions within the sliding window should not be part of a segment or sentence being created.

Abstract: A system and method for creating input data to be used to train a conversational bot may include receiving a set of conversations, each conversation including sentences, classifying each sentence into a dialog act taken from a number of dialog acts, for each set of sentences classified into a dialog act, clustering the set of sentences into clusters based on the content (e.g. text) of the sentences, each cluster having a cluster name or label, and generating a language model based on the cluster labels. Slots may be identified in the sentences based in part on the dialog act classifications. A bot may be trained using data such as the slots, language model, and clusters.

Abstract: Systems and methods of calculating a score for a chain of interactions in a call center, during a first training phase, train a first model which, given an interaction and interaction metadata, predict an initial estimated customer satisfaction score; during a second training phase, train a second model which, given an interaction and interaction metadata, text and metadata of an immediately preceding interaction in a chain of interactions, and features of the chain, predict a refined estimated customer satisfaction score; and during an inference phase: given a chain of interactions and metadata of each interaction, compute an initial estimated customer satisfaction score for each interaction using the first model; beginning with a second interaction in the chain and metadata of each interaction, compute a refined estimated customer satisfaction score for each interaction using the second model; combine the interaction scores into a combined customer satisfaction score; and output the score.

Abstract: A system and method for generating an agent behavioral analytics including transcribing an incoming call to produce a call transcription; and using a trained convolutional neural network (CNN) to produce behavioral labels for the agent in the incoming call for behavioral metrics, based on the call transcription. The CNN may include an embedding layer to convert words in the call transcription into vectors in a word embedding space; a convolution layer to perform a plurality of convolutions on the vectors and to generate vectors of features; a pooling layer to concatenate the vectors of features to a single vector by taking a maximum of each feature generated by the convolution layer; and a classification layer to produce grades of the agent in the incoming call for the set of attributes or behavioral metrics, based on the single vector generated by the pooling layer.

Abstract: Systems and methods of mapping a customer journey in an interactive voice response (IVR) system to a contact reason from a contact reasons list: receive an IVR log comprising a plurality of customer journey entries, wherein each customer journey entry comprises a sequence of one or more menu identifiers; generate an embedding vector for each menu identifier; filter one or more menu identifiers from a menu identifier list, wherein the menu identifier list comprises all menu identifiers contained in the IVR log; cluster one or more remaining menu identifiers from the menu identifier list into one or more clusters, based on the embedding vector of each menu identifier; map each cluster to a contact reason; and create a rule that categorizes a newly received IVR sequence based on a cooccurrence of at least one menu identifier in the newly received IVR sequence and in a given cluster.

Abstract: Systems and methods for mapping a customer journey in an interactive voice response (IVR) system to a category from a categories list, build a directed graph based on one or more sample IVR customer journeys; generate a black list based on the directed graph; for a given customer journey: filter one or more non-informative menus from the plurality of available menus based on the black list; concatenate the respective associated menu prompt of each menu that was not filtered, and one or more user responses to one or more menu prompts from the series of interactions in the given customer journey, into a concatenated word string; calculate a similarity score between the concatenated word string and a category name of each category from the categories list; and map the given customer journey to the category whose category name produces the highest similarity score.

Abstract: A system and method for generating an agent behavioral analytics including transcribing an incoming call to produce a call transcription; and using a trained convolutional neural network (CNN) to produce behavioral labels for the agent in the incoming call for behavioral metrics, based on the call transcription. The CNN may include an embedding layer to convert words in the call transcription into vectors in a word embedding space; a convolution layer to perform a plurality of convolutions on the vectors and to generate vectors of features; a pooling layer to concatenate the vectors of features to a single vector by taking a maximum of each feature generated by the convolution layer; and a classification layer to produce grades of the agent in the incoming call for the set of attributes or behavioral metrics, based on the single vector generated by the pooling layer.

Abstract: Systems and methods of mapping a customer journey in an interactive voice response (IVR) system to a contact reason from a contact reasons list: receive an IVR log comprising a plurality of customer journey entries, wherein each customer journey entry comprises a sequence of one or more menu identifiers; generate an embedding vector for each menu identifier; filter one or more menu identifiers from a menu identifier list, wherein the menu identifier list comprises all menu identifiers contained in the IVR log; cluster one or more remaining menu identifiers from the menu identifier list into one or more clusters, based on the embedding vector of each menu identifier; map each cluster to a contact reason; and create a rule that categorizes a newly received IVR sequence based on a cooccurrence of at least one menu identifier in the newly received IVR sequence and in a given cluster.

Abstract: Systems and methods for mapping a customer journey in an interactive voice response (IVR) system to a category from a categories list, build a directed graph based on one or more sample IVR customer journeys; generate a black list based on the directed graph; for a given customer journey: filter one or more non-informative menus from the plurality of available menus based on the black list; concatenate the respective associated menu prompt of each menu that was not filtered, and one or more user responses to one or more menu prompts from the series of interactions in the given customer journey, into a concatenated word string; calculate a similarity score between the concatenated word string and a category name of each category from the categories list; and map the given customer journey to the category whose category name produces the highest similarity score.

Abstract: A method, system, and non-transitory processor-readable storage medium for automatic key phrase rule generation for automatic key phrase extraction including: receiving a corpus sample including a plurality of documents containing text, receiving a plurality of identified key phrases which relate to a topic of the text of at least one corresponding document; assigning a part-of-speech to each word in the corpus sample; generating a part-of-speech pattern from each identified key phrase; and generating key phrase rules.

Abstract: A system and method for analysis of interactions with a customer service center, comprising receiving a plurality of customer service interactions, receiving a word cloud computed for the plurality of customer service interactions, consolidating similar or synonymous words into word sets, constructing a weighted graph of the word sets, generating a query for the interaction topic based on a corresponding subset of word sets, selecting at least one representative interaction from the retrieved customer service interactions, and displaying at least a portion of the representative interaction for the selected identified topic.